Research is concerned with the development of clinically useful, scientifically dependable methods for assessing interpersonal behavior. Following an exchange conception of social behavior, attention is directed to the analysis of recurrent interaction sequences. Four classes of methods, namely interview, observation, self-report and experimental analogue procedures, are proposed for the study of social interaction. Previous work has led to the preliminary definition of these methods and pilot trials employing them are currently underway. We now propose to employ the methods with 20 married couples who have sought professional help in dealing with marital problems and with 20 couples who claim no serious marital problems. The dependability and other psychometric properties of each method will be established and relationships among methods will be examined. Besides furnishing better means for examining interpersonal behavior, the data we obtain will allow investigation of some basic issues in the taxonomy of attributed interpersonal characteristics, the definition of units of social interaction, and the analysis of nonverbal communication. Sub-projects concerned with these issues are also proposed. The first year of the project period will be devoted to formal data collection. The second year is scheduled mainly for data analysis and for planning research in the modification of disturbed interpersonal behavior toward which our assessment strategies are directed.